test_name: "test GLM-4.7-W8A8C8 PD separation with mooncake layerwise connector"
model: "vllm-ascend/GLM-4.7-W8A8C8"
num_nodes: 2
npu_per_node: 16
env_common: &env_common
  VLLM_USE_MODELSCOPE: true
  OMP_PROC_BIND: false
  HCCL_BUFFSIZE: 512
  HCCL_OP_EXPANSION_MODE: "AIV"
  SERVER_PORT: 8080
  DISAGGREGATED_PREFILL_PROXY_SCRIPT: "examples/disaggregated_prefill_v1/load_balance_proxy_layerwise_server_example.py"

disaggregated_prefill:
  enabled: true
  prefiller_host_index: [0]
  decoder_host_index: [1]

deployment:
  -
    envs:
      <<: *env_common
    server_cmd: >
        vllm serve "vllm-ascend/GLM-4.7-W8A8C8"
        --host 0.0.0.0
        --port $SERVER_PORT
        --data-parallel-size 2
        --data-parallel-size-local 2
        --data-parallel-address $LOCAL_IP
        --data-parallel-rpc-port 13389
        --tensor-parallel-size 8
        --seed 1024
        --enforce-eager
        --enable-expert-parallel
        --max-num-seqs 32
        --max-model-len 8192
        --max-num-batched-tokens 8192
        --trust-remote-code
        --no-enable-prefix-caching
        --gpu-memory-utilization 0.9
        --additional-config '{"enable_shared_expert_dp": true}'
        --kv-transfer-config
        '{"kv_connector": "MooncakeLayerwiseConnector",
        "kv_role": "kv_producer",
        "kv_port": "30000",
        "kv_connector_extra_config": {
                  "prefill": {
                          "dp_size": 2,
                          "tp_size": 8
                  },
                  "decode": {
                          "dp_size": 2,
                          "tp_size": 8
                  }
            }
        }'
  -
    envs:
      <<: *env_common
    server_cmd: >
        vllm serve "vllm-ascend/GLM-4.7-W8A8C8"
        --host 0.0.0.0
        --port $SERVER_PORT
        --data-parallel-size 2
        --data-parallel-size-local 2
        --data-parallel-address $MASTER_IP
        --data-parallel-rpc-port 13389
        --tensor-parallel-size 8
        --seed 1024
        --max-num-seqs 16
        --max-model-len 8192
        --max-num-batched-tokens 8192
        --enable-expert-parallel
        --trust-remote-code
        --no-enable-prefix-caching
        --gpu-memory-utilization 0.9
        --compilation-config '{"cudagraph_mode": "FULL_DECODE_ONLY"}'
        --kv-transfer-config
        '{"kv_connector": "MooncakeLayerwiseConnector",
        "kv_role": "kv_consumer",
        "kv_port": "30200",
        "kv_connector_extra_config": {
                  "prefill": {
                          "dp_size": 2,
                          "tp_size": 8
                  },
                  "decode": {
                          "dp_size": 2,
                          "tp_size": 8
                  }
            }
        }'

benchmarks:
  acc:
    case_type: accuracy
    dataset_path: vllm-ascend/gsm8k-lite
    request_conf: vllm_api_general_chat
    dataset_conf: gsm8k/gsm8k_gen_0_shot_cot_chat_prompt
    max_out_len: 4096
    batch_size: 8
    baseline: 95
    threshold: 10